INTRODUCTION

Epidemiological studies have shown that elevated heart rate(HR) is a marker of coronary artery disease (CAD) progressionand correlates with cardiovascular mortality (1-4). In accordance,experimental data suggest that reduction in resting HR slowsendothelial cell (EC) replication, a marker of endothelialdysfunction, and reduces the progression of atherosclerosis in thecynomolgus monkeys fed a high fat diet (5).

Ivabradine decreases HR by selective inhibition of the sinusnode funny current (If) with no effect on blood pressure orcontractility (6-9), thus is the ideal tool to study the effects of HRreduction on the vessels. Pre-clinical studies have demonstrated aprotective effect of ivabradine on the vasculature (6). In mousemodels of mild (10) or severe dyslipidemia (6, 8, 11-13),

ivabradine prevented endothelial dysfunction and reduced theaortic plaque area (6, 11, 12) and counteracted dyslipidemia-induced reduction of both endothelium-dependent vasorelaxation(8) and aortic distensibility and circumferential cyclic strain (14).In wild type (WT) mice the above cited parameters wereunaffected, either in the presence or absence of ivabradine, eventhough treatment with this drug induced a HR reductioncomparable to the reduction observed in ApoE–/– mice fed a high-fat diet (8). These studies suggest that the protective effect ofivabradine on endothelial function and vascular wall distensibilityis measurable in the presence of an endothelium-damaging agent,such as dyslipidemia, that triggers the onset of atherosclerosis.Ivabradine treatment has been associated with a reduction ofvascular oxidative stress (decrease of NADPH oxidase activity (6,13)), prevention of eNOS uncoupling (13), as well as with a lower

jOURNAL OF PHYSIOLOGY AND PHARMACOLOGY 2018, 69, 1, 35-52www.jpp.krakow.pl | DOI: 10.26402/jpp.2018.1.04

G. AQUILA1, M.B. MORELLI2, F. VIECELI DALLA SEGA1, F. FORTINI1, P. NIGRO3, C. CALICETI4, M. FERRACIN5,M. NEGRINI6, A. PANNUTI7, M. BONORA6, P. PINTON6, R. FERRARI1,8,9, P. RIZZO6,8,9

HEART RATE REDUCTION WITH IVABRADINE IN THE EARLY PHASEOF ATHEROSCLEROSIS IS PROTECTIVE IN THE ENDOTHELIUM

OF ApoE-DEFICIENT MICE

1Department of Medical Sciences, University of Ferrara, Ferrara, Italy; 2IRCCS Neuromed, Angio-Cardio-Neurology Department,Pozzilli, Italy; 3Centro Cardiologico Monzino (IRCCS), Unita di Biologia Vascolare e Medicina Rigenerativa, Milan, Italy;4Department of Chemistry ‘G. Ciamician’, University of Bologna, Bologna, Italy; 5Department of Experimental, Diagnostic

and Specialty Medicine - DIMES, University of Bologna, Bologna, Italy; 6Department of Morphology, Surgery and ExperimentalMedicine, University of Ferrara, Ferrara, Italy; 7Stanley Scott Cancer Center, Louisiana State University Health Sciences Center

and Louisiana Cancer Research Consortium, New Orleans, Louisiana, U.S.A.; 8Laboratory for Technologies of Advanced Therapies(LTTA), University of Ferrara, Ferrara, Italy; 9Maria Cecilia Hospital, GVM Care & Research, E.S. Health Science Foundation,

Cotignola, Italy

Ivabradine, a heart rate reducing agent, protects the vascular system by unidentified mechanisms. We sought todetermine the effects of the treatment with ivabradine, started before plaque formation, on early transcriptional changesand endothelium lesions in regions of aorta subjected to disturbed blood flow. Six week-old apolipoprotein E-deficient(ApoE–/–) mice, fed a low-fat diet, were treated with ivabradine to determine the effect on transcriptional changes (2-and 4-week treatment) and on lesions formation (19-week treatment) in the endothelium of the aortic arch. Microarraysanalysis (60k probes) of endothelium-enriched RNA was carried out to detect changes in gene expression induced bytreatment. Endothelium damage was assessed by en-face immunofluorescence staining for vascular endothelial (VE)cadherin. According to microarray analysis, 930 transcripts were affected by the treatment. We found downregulation ofpro-apoptotic and pro-inflammatory genes, the majority of which are nuclear factor-kB (NF-kB)-and/or angiotensin II-regulated genes, and upregulation of anti-inflammatory genes. Many shear stress-responsive genes were affected by thetreatment and the MAPK, Notch signalling and sterol metabolic processes were among the most significantly affectedpathways. Consistently, we observed increased levels of Hes5, a Notch target gene, together with a reduction ofendothelium damage, in the lower aortic arch of treated- compared with untreated- mice. We concluded that an earlytreatment with ivabradine protected the endothelium of the aortic arch of ApoE–/– mice. Activation of the Notchsignalling could be part of the mechanism underlying this protection.

K e y w o r d s : ivabradine, apolipoprotein E, atherosclerosis, Notch signaling, gene expression, endothelial damage, angiotensinII, shear stress

expression of pro-inflammatory chemokine MCP-1 (6) and ofangiotensin II receptor 1 (AT1-R) (13). Since all the existingstudies have tested the beneficial effects of ivabradine treatment inthe context of advanced stages of atherosclerosis, it is not knownif ivabradine, under dyslipidemic conditions, has an effect onendothelial function before plaque formation.

The mechanisms underlying ivabradine protection are notcompletely understood. It is thought that an improvement ofshear stress, consequent to the HR reduction, is implicated. It iswell known that the ECs exposed to different shear stresspatterns undergo complex transcriptional modifications (15)which modulate their proliferation, survival and activation. Inarterial regions with not-disturbed flow, the ECs express variousatheroprotective genes and suppress several pro-atherogenicones, leading to endothelial stability. In regions with disturbedflow, the atheroprotective genes are suppressed, whereas the pro-atherogenic genes are upregulated, thereby promotingatherosclerosis (16, 17). In vivo (6, 10, 13) and ex vivo studies(10) seem to support this hypothesis and rule out a direct actionof ivabradine on the vessels.

The aim of our study was to investigate the possibility thattreatment with ivabradine, in the early phase of atheroscleroticprocess, before plaque formation, could result in transcriptionalchanges leading to maintenance of normal endothelial function.The study was conducted in 6 weeks old apoliprotein E-deficient(ApoE–/–) mice fed a chow diet (instead of a western diet) andsubjected to early ivabradine treatment.

MATERIALS AND METHODS

Animal treatment

Animal studies were carried out according to the guidelines ofthe European (2010/63/EU) and the Italian (D.L. 26/2014) lawsand after approval by the local ethical review panel of UniversityAnimal House and by the Italian Ministry of University andResearch. Five-weeks old C57BL6/j and ApoE–/– mice (C57/Bl6genetic background), purchased from Charles River Laboratories(Wilmington, MA, US), were exposed to artificial day/night cycle(13 hours light and 11 hours darkness with light on at 7 a.m.) andwere caged, in group of 3, at room temperature (21 – 23°C) with55 – 60% of humidity. Ivabradine (30 mg/kg/day, ivabradinehydrochloride, S 16257-2, Servier, France) was administered indrinking water available ad libitum. Water consumption wasregistered on each water renewal by visual inspection ofivabradine or placebo solution level in the bottles. HR, measuredbefore the beginning of treatment and one week before sacrifice,was monitored by counting the number of waveforms registeredper minute by Doppler echocardiography (Vivid ECG, GEHealthcare Worldwide) and using a pediatric probe (Vividcardiovascular ultrasound 12S R-S, GE Healthcare Worldwide).Animals were not anesthetized during HR monitoring to avoidinterferences with HR values. To reduce stress, which could affectHR, at least 2 days before HR measurement, mice chest wasshaved at third and fourth intercostal space, next to the sternum.Mice were handled carefully, gently held by the neck and placedwith their abdomen up. Nevertheless, this method cannotcompletely prevent stress, which could explain the high HRmeasured at baseline. The calculated HR was the mean of 20consecutive measurements. HR results were expressed as mean ±SEM. Differences between groups were analyzed by one-wayANOVA followed by Dunnett’s test and P < 0.05 was consideredsignificant. For microarray analysis, 6-weeks old ApoE–/– mice (n= 24) were randomly assigned to four groups receiving ivabradineor vehicle for 2 or 4 weeks. For blood sampling and Westernblotting analysis, 6-weeks old ApoE–/– mice (n = 12) were

randomly assigned to two treatment groups receiving ivabradineor vehicle for 4 weeks. For the analysis of endothelium damages,Hes5 expression levels and plaque deposition in the aortic root, 6-weeks old ApoE–/– mice (n = 16) were randomly assigned to twotreatment groups receiving ivabradine or vehicle for 19 weeks.

For blood samples collection mice were sedated with anintraperitoneal injection of Zoletil (60 mg/kg; Parnell Laboratories,Alexandria, NSW, Australia) and Dexdomitor (10 mg/kg; Zoetis,Florham Park, Nj, US). For aortic arch isolation, mice, after beingsedated, were euthanized by an overdose of Zoletil (120 mg/kg;Parnell Laboratories, Alexandria, NSW, Australia).

Endothelial enriched-RNA extraction

In order to increase the specificity of microarrays, RNA wasisolated from endothelium of mice aortic arch only instead ofwhole aorta, adapting the method described by Krenek P et al.,(18). After euthanasia, mice’s hearts were perfused through theleft ventricle with ice-cold saline. When the effluent wascompletely clear, aorta was quickly and carefully isolated andplaced into a Petri dish containing RNAlater solution (RNAstabilization solution, Ambion). Under stereomicroscope(SMZ745T 6x-50x, Nikon, Chiyoda, japan) the adventitial layerand fat portion of the aorta was cleaned off by straining the aortaoppositely with angled forceps. Aortic arch was isolated and amicroloader tip (Eppendorf - Germany) was adapted to aninsulin syringe and filled with 300 µl of Qiazollysis solution(Qiagen, USA). The microloader tip was carefully inserted in theaortic arch until it reached the left common artery, entering fromthe previously removed portion of thoracic aorta. Aortic archwas flushed with Qiazol lysis solution for 30 seconds and theeluate was collected into a RNase-free microfuge tube. As acontrol, RNA was isolated from another whole arch aftermechanical disruption of the tissue using a Douncehomogenizer. In both cases, RNA was extracted using acommercially available kit (miRNeasyMini kit - Qiagen, CA,USA). RNA concentration and purity were determined byAgilent 2100 Bioanalyzer (Agilent Technologies).

Validation of endothelial cells-enriched RNA extraction

Aortic arch of six 10 weeks old C57BL6/j mice were flushedwith different volumes of a phenol containing solution to achieveefficient lysis of ECs only: more than 350 µl of Qiazol lysis buffercaused a disgregation of the whole arch, conversely smallervolume of lysis buffer resulted in insufficient RNA yield (data notshown). Noteworthy, to obtain an efficient RNA extraction and toavoid aorta dissolution, it was needed to constantly perfuse theaorta for not more than thirty seconds. As expected, this methodgave a low yield of RNA (150 – 600 ng from one arch, n = 3–10weeks old C57BL6/j mice) nevertheless every target gene wassuccessfully amplified by qRT-PCR. RNA was also isolated fromthe whole arch (n = 3) after mechanical disruption of the tissueusing a Dounce homogenizer and the same lysis buffer used forflushing the aorta. This method gave good yield and good qualityRNA (1.5 µg total RNA from one arch). Successful endothelialenriched-RNA extraction was confirmed by measuring mRNAlevels of endothelial and smooth muscle specific genes (eNOS andSM22 respectively) (Fig. 1).

Microarray procedures

Total RNA from 12 samples (each containing 70 ng of RNAfrom 2 mice for treatment group) was hybridized on AgilentWhole Mouse Gene Expression Microarray (#G4122F, 60Ktranscripts, Agilent Technologies, Palo Alto, CA, USA). One-color gene expression was performed according to the

36

manufacturer’s procedure. Briefly, labeled cRNA was synthesizedfrom total RNA using the Low RNA Input Linear AmplificationKit (Agilent Technologies, Palo Alto, CA, USA) in the presenceof cyanine 3-CTP (Perkin-Elmer Life Sciences, Boston, MA,USA). Hybridizations were performed at 65°C for 17 hours in arotating oven. Images at 5 µm resolution were generated byAgilent scanner and the Feature Extraction 10.5 software (AgilentTechnologies, Palo Alto, CA, USA) was used to obtain themicroarray raw data. Data transformation was applied to set all thenegative raw values at 1.0 followed by a quantile normalization. Afilter on low gene expression was used to keep only the probesexpressed (Detected) in at least one sample (n = 36, 013).

Microarray data analysis and bioinformatics

Unsupervised principal component analysis (PCA) wasperformed to assess sample similarity and to determine thecontribution of age and treatment to global gene expressionchanges (10K expressed probes, (Qlucore software, Qlucore,Lund, Sweden)). Differentially expressed genes (DEG) wereselected as having a ³ 1.5-fold expression difference betweentreated and untreated groups with a P-value £ 0.01 at moderated t-test (False Discovery Rate, FDR, 7%). Hierarchical clustering wasperformed with GeneSpring Clustering tool (Agilent Technologies,Palo Alto, CA, USA) using the list of DEG and the Pearsoncentered correlation as a measure of similarity between samples.Correlations between continuous variables were tested by Pearsonanalysis. The Database for Annotation, Visualization andIntegrated Discovery (DAVID) (19) was used to perform GeneOntology (GO) functional and Kyoto Encyclopedia of Genes andGenomes (KEGG) and Biocarta pathways enrichment analyses(threshold P-value < 0.1 and enrichment gene count > 2).

Reverse transcription and real time PCR

Total RNA from 6 samples (each containing 70 ng of RNAfrom 2 mice for treatment group) was reverse transcribed in avolume of 25 µl using 250 units of SuperScript III reversetranscriptase and 50 ng of random hexamers. 2 µl of the cDNAmixture were used for real-time qPCR. Real-time qPCRreactions (final volume of 25 µl) were performed on anApplied Biosystems 7500 Fast Real-Time PCR System usingPerfeCta SYBR Green SuperMix with ROX kit (QuantaBiosciences, Gaithersburg MD, USA) according to themanufacturer’s protocol. Changes in gene expression werecalculated by the 2–DDCt formula using Rpl13a as referencegene. The following primers were used:

RPL13: forward 5’-AGCCCAGGGTGCTTTGCGG-3’,reverse 5’-GCGCCATGGCTGCCTCCTATAC-3’;

SM22: forward: 5’-GGGCGGCAGAGGGGTGACAT-3’,reverse: 5’-TGAGGCAGAGAAGGCTTGGTCGT-3’;

NPPC: forward 5’-ACACCACCGAAGGTCCCG-3’,reverse: 5’-TCGGTCTCCCTTGAGATTGG-3’;

OLR1: forward 5’-TGCAAACTTTTCAGGTCCTTGT-3’,reverse: 5’-AACTGGCCACCCAAAGATTG-3’;

eNOS: forward: 5’-TTCCCCGCCTAGTCCTCGCC-3’,reverse: 5’-CCGGGGGTCCTGGCTGAGAG-3’.

Results were expressed as mean ± SEM. Differencesbetween groups were analyzed by unpaired t-test, P-value £ 0.05was considered significant.

Western blotting

Immediately after euthanasia, the mouse aorta was perfusedwith saline solution then excised, cleaned of fat and fibrousmaterial in a Petri dish containing ice cold saline solution.Segments of aortic arch were pooled (n = 6 per treatment group)and immersed in 0.5 ml of RIPA buffer (0.05% sodiumdeoxycholate was freshly added) containing 10 µg/ml ofaprotinin, 10 µg/ml of leupeptin, 10 µg/ml of pepstatin A, 1 mMPMSF, and 1 mM sodium orthovanadate. Samples werehomogenized by Dounce homogenizer on ice for 30 min. Proteinconcentration was quantified by by Pierce BCA Protein AssayKit (Thermo Scientific, Wilmington, DE). Immunoblotting wasperformed as described in (20, 21), using the following primaryantibodies: rabbit anti-OLR-1 (Abcam, Cambridge, UK, 1:1000,Cat. ab60178), rat anti-VE-cadherin (BD Pharmigen, San Diego,CA, USA, 1:500, Cat. 555289) and rabbit anti-a-smooth muscleactin (a-SMA) (Novus Biologicals, Littleton, CO, USA, 1:500,Cat. NB600).

En-face analysis of vascular endothelial-cadherin and Hes5staining in the endothelium of the lesser curvature of the aortic arch

Immediately after mouse euthanasia, the aorta was perfusedwith saline solution and fixed for 10 minutes with 4%paraformaldehyde. Aorta was then excised, cleaned of fat andfibrous material in a Petri dish containing ice cold saline solution.Segments of aortic arch corresponding to lesser curvature (22)were immersed in 1 ml of blocking buffer (PBS 1X, 0.1% TritonX-100, 2% BSA) and left rocking at room temperature for 1.5hours. Primary antibody incubation (rat anti-VE-cadherin, BDPharmigen, San Diego, CA, USA, 1:100, Cat. 555289; rabbit anti-

37

Ct (

SM22

- eN

OS) ****

Fig. 1. Enrichment of endothelial specific markers inRNA extracted by using the ‘flushing technique’.Enrichment in eNOS and SM22 mRNA levels is shownby the difference in the relative Ct numbers between theRNA extracted from whole aortic arch (W-AORTA, n =3) and RNA extracted by the ‘flushing’ technique (E-AORTA, n = 3). Differences between groups wereanalyzed by unpaired t-test and P < 0.05 was consideredsignificant (****P < 0.0001).

Hes5, Santa Cruz Biotechnology, Santa Cruz, CA, USA, 1:100,Cat. sc-25395) was performed overnight at 4°C. After washing theaortic segments three times in washing buffer (PBS 1X, 0.1%Triton X-100), secondary antibody, previously diluted in blocking

buffer, was added and incubated for 1.5 hours at 4°C. Thesecondary antibody used were Alexa Fluor 546 Goat Anti-Rat IgG(Life Technologies, CA, USA, 1:500, Cat. A11081) and AlexaFluor 633 Goat Anti-Rabbit IgG (Life Technologies, CA, USA,

38

********

Fig. 2. Heart rate reduction induced by ivabradinetreatment. Heart rate (BPM) in ApoE–/– mice beforeand after two or four weeks of ivabradine treatment.Results are expressed as mean ± SEM. Differencesbetween groups were analyzed by one-way ANOVAfollowed by Dunnett’s test and P < 0.05 wasconsidered significant (****P-value < 0.0001).

Fig. 3. Ivabradine treatment does not alter lipid profile and body weight of ApoE–/– mice. Serum levels of (A) LDL/VLDL and (B)HDL in ApoE–/– mice before and after four weeks of ivabradine treatment. Results are expressed as mean ± SEM. Differences betweentreatment groups were analyzed by unpaired t-test and P < 0.05 was considered significant. (C) Effect of ivabradine treatment on bodyweight in ApoE–/– mice. Differences between treated and untreated mice, in each age group, were analyzed by unpaired t-test and P <0.05 was considered significant.

1:500, Cat. A21070). After washing three times, aortic specimenswere placed on a glass slide with the intima side up and, beforecovering samples with a coverslip, a drop of mounting media wasapplied directly to fluorescently labeled tissue samples (ProLongAntifade Reagents, Life Technologies, CA, USA). Five images ofthe endothelial monolayer were obtained using a Zeiss LSM 510confocal microscope (Zeiss, jena, Germany, 40 × magnification).Lack of VE-cadherin staining was the criteria used to identify theareas with damaged endothelium. Hes5 protein levels and thedamaged endothelium area were calculated by using Image jsoftware (Image j analysis software - http://imagej.nih.gov/ij/).Results were expressed as mean ± SEM. Differences betweengroups were analyzed by unpaired t-test and P < 0.05 wasconsidered significant.

Plaque size analysis in section of aortic root

Hearts of 25 weeks old ApoE–/– mice, following 19 weekstreatment with ivabradine (30 mg/kg/day, started at 6 weeks ofage) or vehicle, were snap-frozen and samples were cut on aLeica (Wetzlar, Germany) cryostat into 10 µm sections, startingat the apex until was reached the aortic valve area. At least fourconsecutive sections were fixed in Baker’s fixative, stained inOil Red O and hematoxylin/eosin solution and mounted on glassslides. Images were acquired by Aperio microscope (Leica,Wetzlar, Germany). Percentage of cross-sectional aortic valve

area occupied by plaque was quantified by using Image jsoftware (Image j analysis software - http://imagej.nih.gov/ij/).Results were expressed as mean ± SEM. Differences betweengroups were analyzed by unpaired t-test and P < 0.05 wasconsidered significant.

Serum analysis

Blood samples were collected from mice tail. Serum wasprepared and stored at –80°C. A commercially available ELISAkit was used to measure levels of HDL and LDL/VLDLcholesterol (Cat. Ab 65390, Applied Biosystems, Waltham, MA,USA). Results were expressed as mean ± SEM. Differencesbetween groups were analyzed by unpaired t-test and P < 0.05was considered significant.

RESULTS

Heart rate, body weight and lipid levels

We chose to start the treatment in 6 week-old ApoE–/– mice asthis age corresponds to the early stage atherosclerosisdevelopment, i.e. before monocyte adhesion (23). For microarrayanalysis, treatment was administered for 2 and 4 weeks and HRreduction was confirmed for each mouse in both treatments groups

39

Fig. 4. Unsupervised principal components analysis. Unsupervised principal components analysis (10K expressed sequences).Principal component analysis (PCA) plot was derived from transcriptome microarray data of treated and placebo mice, analyzed after2 and 4 weeks of ivabradine treatment in 8 and 10 week-old ApoE–/– mice. The PCA axis reflect values derived from orthogonal linearcombination of the variables in the data set (128). The percentage of variance accounted by each principal component is reported inthe x-, y- and z-axes labels. As shown, the clustering of samples depended mainly on presence/absence of ivabradine treatment (bluespheres for treated mice and yellow spheres for untreated mice), rather than age or treatment duration. Blue and yellow spheres indicatetreated and untreated mice, respectively. Circle and triangles indicate 2- and 4-week treatment, respectively.

(data not shown). HR decreased from 751.5 ± 4.8 bpm to 604 ±12.7 bpm (19.6%) after 2 weeks treatment (P < 0.0001) and to 550± 15 bpm (27%) after 4 weeks (P < 0.0001) (Fig. 2). The dose of30 mg/kg/day of ivabradine was chosen as in preliminary dosefinding experiments, it ensured a reduction of HR in the desiredrange of 13 – 23% (6, 14). Similar HR reduction was obtained in 6week-old ApoE–/– mice, treated with ivabradine for 4 weeks, usedfor protein levels assessment, and 19 weeks, used for endotheliumfunction assessment (data not shown). Ivabradine had not effect onbody weight after 2 and 4 weeks treatment (Fig. 3A) and on serumlevels of LDL/VLDL (110.6 ± 18 mg/dl and 140 ± 19 mg/dl,control versus 4 weeks treatment group) and HDL (44 ± 23 mg/dland 40 ± 20 mg/dl, control versus 4 weeks treatment group) (Fig.3B). Lastly, during the 19 weeks of ivabradine treatment there wasno statistically significant differences in body weight betweenuntreated and treated animals (data not shown).

Identification of differentially expressed genes in endothelium

The gene expression profile of treated or placebo mice wasanalyzed after 2 or 4 weeks treatment in 8 and 10 week-old

ApoE–/– mice, respectively. Unsupervised PCA showed that mostof the variation in gene expression was due to ivabradinetreatment and it was independent of treatment duration and ageof the mice (Fig. 4). Based on the PCA results, a list ofdifferentially expressed genes (DEG) was obtained bycomparing the gene expression profile between the pool oftreated (2- or 4- weeks treatment, 2T and 4T, respectively) anduntreated mice (2NT and 4NT). The treatment altered theexpression of 930 transcripts, 630 were up-regulated and 300down-regulated (P-value cut-off: 0.01, fold change ³ 1.5).Among the modulated sequences we identified 73uncharacterized expressed sequence tags (ESTs), 61 RikencDNA sequences and 314 long intergenic non-coding RNAs(lincRNA). Fig. 5 shows that the heatmap of hierarchicalclustering of gene expression in samples from untreated andivabradine-treated mice. To exclude that the differences in geneexpression between groups were due to a different content ofvascular smooth muscle cell (VSMCs) mRNA, we performed aclustering analyses for selected VSMCs-specific genes (Acta2,Tagln and Cnn1). These analyses showed that treated anduntreated groups were not separated when the specific VSMCs

40

Fig. 5. Heatmap ofhierarchical clustering ofgene expression insamples from untreatedand ivabradine-treatedmice. Heatmap ofuntreated and ivabradine-treated mice (each groupcontains mice treated for2 and 4 weeks) obtainedusing 930 differentiallyexpressed genes. Therepresented genes (rows)and the experimentalgroups (columns) weresorted out by hierarchicalclustering. High- andlow-expression is referredto the treated group.

transcripts were considered (Fig. 6A). In addition, the threeVSMCs-specific gene expression values co-correlated (Tagln,Cnn1, R2 = 0.93, P < 0.0001 and Cnn1, Acta2, R2 = 0.902, P <0.0001) (Fig. 6B and 6C), while there was no significantcorrelation between the expression levels of ivabradine-modulated transcripts and VSMCs-specific transcripts (i.e.Rasgrf2, Cnn1, R2 = 0.0119) (Fig. 6D).

Therefore, the analysis strongly suggests that the differencesin gene expression profile between treated and untreated micecannot be explained by the different contribution of VSMCstranscripts.

Gene ontology functional and pathway enrichment analysis ofdifferentially expressed genes

A total of 140 gene ontology (GO) biological processes wereenriched by ivabradine treatment. The most significantlybiological process affected were those related to sterolmetabolism (P = 0.000097), cholesterol metabolic process (P =0.00034), epithelial cells differentiation (P = 0.00053) andepithelium development (P = 0.00063) (Table 1). The KEGGSand BIOCARTA pathways affected by treatment are shown inTable 2. The MAPK signalling pathway and the steroidbiosynthesis (P = 0.0013 and P = 0.0032, respectively), as wellas the Notch signalling pathway (P = 0.009) were among themost significantly regulated. Tables 3 and 4 show individualgenes belonging to the above cited pathways and the fold-changes in their expression following ivabradine treatment.

Ivabradine modulates expression of genes regulating epithelialcells functions and inflammation

To understand the possible functional significance of thegenes affected by the treatment in the atherosclerotic process, weperformed a literature search which showed that an elevatednumber of genes linked to ECs dysfunctions and/orinflammation are present among the DEG (Table 5). Morespecifically, ivabradine downregulated the expression of pro-inflammatory (Calca, Calcb, Tnfsf, Has1, Tslp, Ptgs2, Wnt7A)and pro-apoptotic (Serpinb5, Olr1) genes, whereas anti-inflammatory (Ers1, Ers2, Ffar3) and anti-apototic genes(Park2, Pax2, Dad1) were upregulated. Furthermore, genesassociated to increased endothelium permeability (Prkca, Ldlr1,Adamts7, Adamts8) were downregulated, whereas genesinvolved in endothelium repair (Hes5, Terc) were induced bytreatment. Nppc and Pdzk1, both coding for proteins secreted bythe endothelium to limit damages induced by inflammatoryconditions, were among the genes more affected by ivabradinetreatment. Quantitative RT-PCR confirmed the down-regulationof Olr1 and Nppc following 4 weeks treatment with ivabradine(Fig. 7A and 7B). Additionally, Western blot analysis showedthat ivabradine treatment also downregulated OLR-1 protein(Fig. 7C).

The analysis of the genes involved in cholesterol metabolicprocess which resulted to be modulated by ivabradine showeddownregulation of the low density lipoprotein receptor gene(Ldlr) and of several other genes involved in cholesterol

41

Fig. 6. Clustering analyses for selected VSMCs-specific genes. (A) Heatmap Acta2, Tagln and Cnn1 genes in samples from micetreated with ivabradine for 2 and 4 weeks or untreated. The random distribution of treated and untreated groups is highlighted in redbox. (T, treated; NT, untreated); (B) Correlation analysis between expression levels of two ivabradine-modulated genes (Rasgrf2;Prkca); (C) Correlation analysis between expression levels of two VSMCs transcripts genes (Tagln; Cnn1); (D) Correlation analysisbetween expression levels of one of the ivabradine-modulated gene (Rasgrf2) and a VSMCs transcript (Cnn1). (2 T and 4 T, 2 week -4 week treatment; 2 NT and 4 NT, control groups).

metabolism (Hmgcs1, Nsdhl, Cyp51, Ch25h, Sc4mol) andupregulation of Hnf1A and Cyp46a1, which are involved in theregulation of HDL cholesterol levels and cholesteroldegradation (Table 6). The genes reported in Table 6, are alsoinvolved in atherosclerosis onset and progression. For instance,Ldlr, downregulated by ivabradine treatment, induces uptake ofLDL in endothelial cells leading to alteration of lipid dynamicin cell membrane and increased monocytes adherence (24).Ch25h, downregulated by the treatment, plays a role ininflammation, in addition to the more established role in theregulation of cholesterol homeostasis (25). This gene encodesfor an enzyme which catalyzes the formation of 25-hydroxycholesterol from cholesterol, which stimulates

macrophage foam cell formation in Apoe–/– mice (26). Sc4mol,Cyp51, Hmgcs1 and Ldlr are target of SREBP1 (sterolregulatory element-binding proteins), a family of transcriptionfactors transiently induced by laminar flow but continuouslyunpregulated under disturbed flow. Their downregulation couldbe an indication of ivabradine-mediated modulation of shearstress and reduced disturbed flow (27-29). Finally, Lepr(upregulated by ivabradine treatment) seems to conferprotection against atherosclerosis, since mice defective in allleptin receptor signaling pathways develop severe obesity,hypercholesterolemia, and increased atherosclerosis (30).

Pubmed search also showed that many genes involved ininflammation and endothelial dysfunction, which resulted to be

42

GO biological process Count P-value Fold Enrichment

sterol metabolic process 9 0.000097 6.2cholesterol metabolic process 8 0.00034 6epithelial cell differentiation 10 0.00053 4.3epithelium development 15 0.00063 2.9developmental maturation 9 0.00064 4.7reproductive structure development 10 0.00079 4.1steroid metabolic process 11 0.00095 3.6reproductive developmental process 14 0.0015 2.8prostate gland morphogenesis 5 0.0015 9.8tissue morphogenesis 13 0.0019 2.9

Count: the number of DEGs involved in the enriched biological process (> 2). P-value < 0.1.

Table 1. Gene ontology (GO) functional enrichment analysis of differentially expressed genes (DEGs) - top ten.

Category Term Count P-valueKEGG_PATHWAY Notch signalling pathway 5 0.009KEGG_PATHWAY MAPK signalling pathway 11 0.013KEGG_PATHWAY Steroid biosynthesis 3 0.032BIOCARTA Segmentation clock 3 0.047

BIOCARTA Downregulated of MTA-3in ER-negative breast tumors 3 0.047

BIOCARTA Presenilin action in Notch and Wnt signalling 3 0.054

KEGG_PATHWAY Basal cell carcinoma 4 0.064KEGG_PATHWAY Tight junction 6 0.075KEGG_PATHWAY Melanogenesis 5 0.086

Count: the number of DEGs involved in the pathways (>2). P-value < 0.1.

Table 2. The enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) and Biocarta pathways for differentially expressed genes(DEGs).

Gene Name Gene Symbol P-value FC (Abs) Regulation

hairy and enhancer of split 5(Drosophila) Hes5 0.0077 3.34 up

jagged 2 Jag2 0.0028 2.21 downNotch gene homolog 1(Drosophila) Notch1 0.0037 1.59 up

recombination signal binding proteinfor immunoglobulin kappa J region Rbp-jk 0.0045 1.53 up

mastermind like 3(Drosophila) Maml3 0.0091 1.52 up

Regulation: type of modulation following ivabradine treatment. P-value < 0.1. Abbreviation: FC (Abs), absolute fold-change.

Table 3. Genes of the mitogen-activated protein kinases (MAPK) signalling pathway regulated by ivabradine.

43

Gene Name Gene Symbol P-value FC (Abs) Regulation

hairy and enhancer of split 5(Drosophila) Hes5 0.0077 3.34 up

jagged 2 Jag2 0.0028 2.21 downNotch gene homolog 1(Drosophila) Notch1 0.0037 1.59 up

recombination signal binding proteinfor immunoglobulin kappa J region Rbp-jk 0.0045 1.53 up

mastermind like 3(Drosophila) Maml3 0.0091 1.52 up

Regulation: Type of modulation following ivabradine treatment. P-value < 0.1. Abbreviations: FC (Abs), absolute fold-change.

Table 4. Genes of the Notch signalling pathway regulated by ivabradine.

Gene Name Gene Symbol FC (Abs) P-value Regulation Function Ref(s)natriuretic peptide type C Nppc 5.73 0.0018 down marker endothelial

damage(61, 62)

PDZ domain containing 1 Pdzk1 5.20 0.0002 down marker of

endothelial damage(63, 64)

wingless-related MMTV integration site 7A Wnt7a 5.17 0.0079 down marker of

inflammation (65)

serine (or cysteine) peptidase inhibitor, clade B, member 5 Serpinb5 5.02 0.0013 down pro-apoptotic (66)

a disintegrin-like and metallopeptidase with thrombospondin type 1 motif, 8 Adamts8 4.67 0.0068 down increases

permeability (67)

gap junction protein, beta 3 Gjb3 4.40 0.0065 down pro-inflammatory (68)calcitonin-related polypeptide, beta Calcb 3.58 0.0047 down pro-inflammatory (69)calcitonin/calcitonin-related polypeptide, alpha Calca 3.49 0.0012 down pro-inflammatory (70)

solute carrier family 6 (neurotransmitter transporter, serotonin), member 4

Slc6a4 3.42 0.0002 down Pro-inflammatory (71)

hairy and enhancer of split 5 (Drosophila) Hes5 3.34 0.0077 up endothelium

integrity (40)

oxidized low density lipoprotein (lectin-like) receptor 1 Olr1 3.19 0.0036 down pro-apoptotic (72)

a disintegrin-like and metallopeptidase with thrombospondin type 1motif, 7 Adamts7 2.75 0.0042 down increases

permeability (73)

paired box gene 2 Pax2 2.44 0.0005 up anti-apoptotic (74)tumor necrosis factor (ligand) superfamily, member 18 Tnfsf18 2.25 0.0092 down pro-inflammatory (75)

hyaluronan synthase 1 Has1 2.23 0.0063 down pro-inflammatory (76)

low density lipoprotein receptor Ldlr 2.14 0.0052 down increases permeability (27)

thymic stromal lymphopoietin Tslp 1.99 0.0026 down pro-inflammatory (77)Esr1 1.97 0.0048 up anti-inflammatory (78)

aTP-binding cassette, sub-family B (MDR TAP), memeber 1A Abcb1a 1.93 0.0028 down cholesterol efflux (79)

prostaglandin-endoperoxide synthase 2 Ptgs2 1.91 0.0060 down pro-inflammatory (80)

telomerase RNA component Terc 1.87 0.0066 up endothelium integrity (81)

cholesterol 25-hydroxylase Ch25h 1.82 0.0021 down pro-inflammatory (26)CD55 antigen Cd55 1.77 0.0092 down pro- inflammatory (82)free fatty acid receptor 3 Ffar3 1.61 0.0034 up anti-inflammatory (83)

Esr2 1.55 0.0097 up anti-inflammatory (78)paraxonase 2 Pon2 1.55 0.0021 up anti-inflammatory (84)parkinson disease (autosomal recessive, juvenile) 2, parkin Park2 1.52 0.0052 up anti- apoptotic (85)

Prkca 1.52 0.0047 down Increases permeability (86)

Regulation: Type of modulation following ivabradine treatment. P-value < 0.01; FC(Abs) > 1.5. Abbreviations: FC (Abs), absolutefold-change; Ref(s), references.

Table 5. Genes associated to endothelium dysfunctions and inflammation regulated by ivabradine.

downregulated by ivabradine, are targets of the nuclear factor-kappaB (NF-kB) (Table 7) and angiotensin II (AngII) signalling

pathways (Table 8). Surprisingly, ivabradine modified theexpression of these genes but not the expression of genes

44

Fig. 7. Ivabradine-mediated down-regulation of Olr1 and Nppc in 10week-old ApoE–/– mice. qRT-PCRanalysis of Olr1 (A) and Nppc (B)expression in 10 week-old ApoE–/–

mice, following 4 weeks oftreatment with ivabradine (30mg/kg/day in drinking water, n = 6)or no treatment (n = 6). Relativechanges in mRNA expression levelswere calculated according to the 2–

DDCt method using RPL13 asreference gene. Results areexpressed as mean ± SEM (n = 3,each sample is a pool of two ApoE–

/– mice aortic arch ECs enriched-RNA). Differences between groupswere analyzed by unpaired t-test andP < 0.05 was considered significant(****P < 0.0001). T, treated; NT,untreated. (C) Western blot analysisof OLR-1 in 10 week-old ApoE–/–

mice, following 4 weeks oftreatment with ivabradine (30mg/kg/day in drinking water, n = 6,4T) or no treatment (n = 6, 4NT).VE-cadherin and a-SMA were usedfor equal protein loading.

Gene Name Gene Symbol P-value FC

(Abs) Regulation Function Ref(s)

similar to Hmgcs1 protein; 3-hydroxy-3-methylglutaryl-Coenzyme A synthase 1

Hmgcs1 0.0004 3.29 down Cholesterol biosynthesis (87)

cytochrome P450, family 51 Cyp51 0.0034 2.27 down Cholesterol biosynthesis (88)

low density lipoprotein receptor Ldlr 0.0052 2.14 down Cholesterol homeostasis (27)

HNF1 homeobox A Hnf1a 0.0093 2.08 upRegulator of HDL-

cholesterol metabolism

(89)

leptin receptor Lepr 0.0083 1.96 up Cholesterol metabolism (90)

NAD(P) dependent steroid dehydrogenase-like Nsdhl 0.0054 1.94 down Cholesterol

biosynthesis (90)

cytochrome P450, family 46, subfamily a, polypeptide 1 Cyp46a1 0.0094 1.87 up Cholesterol

homeostasis (92)

cholesterol 25-hydroxylase Ch25h 0.0022 1.82 down Cholesterol metabolism (93)

sterol-C4-methyl oxidase-like Sc4mol 0.0072 1.62 down Cholesterol biosynthesis (87)

Regulation: type of modulation following ivabradine treatment. P-value < 0.01; FC(Abs) > 1.5. Abbreviations: FC (Abs), absolutefold-change; Ref(s), references.

Table 6. Genes of the cholesterol metabolic process affected by ivabradine.

coding for NF-kB transcription factors (Rela, Relb, Nfkb1,Nfkb2) and for angiotensin II receptor 1 (Agtr1) or angiotensinI converting enzyme (Ace). Interestingly, Table 9 shows thatseveral genes known to be shear stress-regulated were affectedby treatment (31).

Ivabradine induces Notch signalling, reduces atheroscleroticlesion formation in the aortic root and limits endothelialdamage in ApoE–/– mice aortic arch

Confocal microscopy-aided en face analysis of the VE-cadherin staining in the lesser curvature of aortic arch showeda difference in the percentage of area characterized bydamaged endothelium in favour of treated mice (91.74 ± 4.084

% versus 28.98 ± 7.279 %, untreated and treated, respectively,P < 0.0001) (Fig. 8A). We also performed en face analysis ofHES5 expression levels in the areas of the endothelium ofuntreated mice still partially intact (about 10% of the analysedarea). The analysis showed that treatment induced HES5, amarker of Notch signalling activation, in the endothelium ofthe lesser curvature (4.1 × 106 ± 0.3 × 106 AFU versus 6.2 ×106 ± 0.6 × 106 AFU, untreated and treated, respectively, P <0.05) (Fig. 8B). Oil Red O staining of sections of aortic rootshowed that the atherosclerotic plaque area was reduced by44% in the aortic root of ivabradine-treated ApoE–/– micecompared with untreated ApoE–/– mice (18.81 ± 0.6% versus10.48 ± 0.97 %, untreated and treated, respectively, P <0.0001) (Fig. 9).

45

Gene name Gene Symbol

FC(abs)

Ivabradine Regulation

NF- BRegulation Ref(s)

PDZ domain containing 1 Pdzk1 5.20 down up* (97)(94)(93)(92)

Wingless-related MMTV integration site 7A Wnt7a 5.18 down up (95)Serine (or cysteine) peptidase inhibitor, clade B, member 5 Serpinb5 5.03 down up (96)

Gap junction protein, beta 3 Gjb3 4.40 down up* (97)Calcitonin-related polypeptide, beta Calcb 3.58 down up (98)Solute carrier family 6 (neurotransmitter transporter, serotonin), member 4 Slc6a4 3.42 down up* (99)

Oxidized low density lipoprotein (lectin-like) receptor 1 Olr1 3.20 down up* (100)

Interleukin 11 Il11 2.83 down up* (101)A disintegrin-like and metallopeptidase with thrombospondin type 1 motif, 7 Adamts7 2.74 down up (102)

Hyaluronan synthase1 Has1 2.23 down up (103)Prostaglandin-endoperoxide synthase 2 Ptgs2 1.91 down up (104)CD55 antigen Cd55 1.77 down up (105)

Regulation: Type of modulation following ivabradine treatment. P-value < 0.01; FC(Abs) > 1.5; * indicated that the gene has an NF-kB site in the promoter but has not clearly been shown to be controlled by NF-kB; or the gene expression is associated with increasedNF-kB activity but has not been shown to be a target directly. Abbreviations: FC (Abs), absolute fold-change; NF-kB, nuclear factor-kappa B; Ref(s), references.

Table 7. Genes target of NF-kB regulated by ivabradine.

Gene Name Gene Symbol FC(abs)

Ivabradine Regulation

Angiotensin II Regulation Ref(s)

T-cell leukemia, homeobox 2 Tlx2 7.09 down up (106)oxidized low density lipoprotein (lectin-like) receptor 1 Olr1 3.20 down up (107)

low density lipoprotein receptor Ldlr 2.14 down up (108)muscle and microspikes RAS Mras 2.11 down up (109)thymic stromal lymphopoietin Tslp 2.00 down up (110)prostaglandin-endoperoxide synthase 2 Ptgs2 1.91 down up (111)regulator of G-protein signalling 4 Rgs4 1.83 up down (112)phosphodiesterase 3A, cgmp inhibited Pde3a 1.60 up down (113)regulator of calcineurin 1 Rcan1 1.53 down up (114)

Regulation: type of modulation following ivabradine treatment. P-value < 0.01; FC(Abs) > 1.5; Abbreviations: FC (Abs), absolutefold-change; AngII, Angiotensin II; Ref(s), references. *indicates that the gene is a regulator of the AngII pathway.

Table 8. Angiotensin II pathway related genes modulated by ivabradine.

46

Gene Name Gene Symbol FC(Abs) P-value Regulation Ref(s)

natriuretic peptide type C Nppc 5.73 0.0018 down (115)gap junction protein, beta 3 Gjb3 4.40 0.0065 down (116)oxidized low density lipoprotein (lectin-like) receptor 1 Olr1 3.19 0.0036 down (117)

a disintegrin-like and metallopeptidase with thrombospondin type 1motif, 7 Adamts7 2.75 0.0042 down (73)

jagged 2 Jag2 2.22 0.0028 down (118)low density lipoprotein receptor Ldlr 2.14 0.0052 down (27) muscle and microspikes RAS Mras 2.11 0.0064 down (119)ATP-binding cassette, sub-family B (MDR TAP), memeber 1A Abcb1a 1.93 0.0028 down (120)

prostaglandin-endoperoxide synthase 2 Ptgs2 1.91 0.0060 down (121)regulator of G-protein signalling 4 Rgs4 1.83 0.0089 up (122)mitogen-activated protein kinase kinasekinase 14 Map3k14 1.70 0.0013 up (123)dual specificity phosphatase 14 Dusp14 1.64 0.0090 down (124)sterol-C4-methyl oxidase-like Sc4mol 1.61 0.0072 down (125)Notch gene homolog 1 Notch1 1.59 0.0037 up (42)related RAS viral (r-ras) oncogene homolog 2 Rras2 1.59 0.0024 down (124)protein kinase C, alpha Prkca 1.52 0.0046 down (126)mitogen-activated protein kinase kinasekinase 8 Map3k8 1.52 0.0030 down (127)

Regulation: type of modulation following ivabradine treatment. P-value < 0.01; FC(Abs) > 1.5. Abbreviations: FC (Abs), absolutefold-change; Ref(s), references.

Table 9. Shear stress sensitive genes modulated by ivabradine.

Fig. 8. Notch signalling modulation and endothelial damage reduction by ivabradine treatment in ApoE–/– mice aortic arch. (A)Representative staining for VE-cadherin (red) in 25 week-old ApoE–/– mice aortic arch lesser curvature, following (2) 19 weeks oftreatment with ivabradine (30 mg/kg/day in drinking water) or (1) no treatment (40 × magnification). (3) Difference in area withendothelium damages between untreated (n = 8) and treated (n = 8) mice. Results are expressed as mean ± SEM. Differences betweengroups were analyzed by unpaired t-test and P < 0.05 was considered significant (***P < 0.001). (B) Representative staining for VE-cadherin (red) and Hes5 (green) in 25 week-old ApoE–/– mice aortic arch lesser curvature, following (2) 19 weeks of treatment withivabradine (30 mg/kg/day in drinking water) or (1) no treatment (40 × magnification). (3) Difference in Hes5 protein level betweenuntreated (n = 8) and treated (n = 8) mice. Results are expressed as mean ± SEM. Differences between groups were analyzed byunpaired t-test and P < 0.05 was considered significant (*P < 0.05).

DISCUSSION

The main finding of this study is that early HR reduction withivabradine induces an atheroprotective gene expression profile inthe aortic arch endothelium of ApoE–/– mice fed a low fat diet.

The atheroprotective milieu induced by ivabradine consistsin an enhancement of anti-apoptotic, anti-inflammatory and pro-endothelium repair genes with a concomitant decrease ofexpression of genes that increase endothelium permeability andfavor inflammation and apoptosis. These effects of earlytreatment are linked with maintenance of endothelial integrity

and reduction in plaque area in the aortic root, detected at a laterstage (after 19 weeks). These results are in agreement with thoseof other authors which have shown a reduction of lesions in theaorta of older and dyslipidemic mice treated with ivabradine (6).

Our data also show, for the first time, that early treatmentwith ivabradine upregulated Hes5, Notch1, Maml1, Rbp-jk,suggestive of the activation of the Notch signalling pathway.This hypothesis is reinforced by the downregulation induced byivabradine of Jagged2, coding for a Notch ligand. The Notchligand jagged1 (a member of the jagged/Serrate family to whichbelongs jagged2) is a weaker activator of endothelial Notch

47

Fig. 9. Plaque area reduction inaortic root by ivabradine treatmentin ApoE–/– mice aortic arch.Hematoxylin and Oil Red stainingin 25 weeks old ApoE–/– mice aorticvalves following treatment with (1)vehicle or (2) ivabradine (30mg/kg/day, started at 6 weeks ofage). Both groups were fedstandard diet. (3) Difference in %aortic root plaque area betweenuntreated (n =4) and treated (n = 4)mice. Results are expressed asmean ± SEM. Differences betweengroups were analyzed by unpairedt-test and P < 0.05 was consideredsignificant (*****P < 0.0001).

Fig. 10. Proposed model for ivabradine-mediated atheroprotection in the endothelium of ApoE–/– mice lower aortic arch.AngII/ERK1/2-mediated activation of NF-kB induces transcription of pro-atherosclerotic downstream genes (highlighted in red).Ivabradine treatment, leads to Notch activation which in turn interferes with NF-kB activity by inhibiting ERK1/2 activation.

signalling in comparison to Delta-like ligand 4 (32). Thereduction of Jagged2 by ivabradine could lead to increasedDelta-like ligand 4 mediated-signalling and, therefore, toactivation of the Notch pathway. Changes in expression levels ofNanog and Tlx2, Notch target genes in breast cancer (33) andneuronal cells (34) respectively, are also suggestive of anactivation of Notch pathway by ivabradine treatment. Alsoconsistent with the activation of Notch signalling, Hes5 proteinwas found upregulated in the endothelium of the lower aorticarch of ApoE–/– treated with ivabradine for 19 weeks.

During the last 10 years, the role of Notch in theendothelium has been actively investigated. Except for twostudies showing a proatherosclerotic role of Notch (35, 36), thereis growing evidence for a protective role of this pathway in theendothelium (37-41). To this end, it is relevant that laminar shearstress-modulated transcription of genes belonging to the Notchpathway has been found in ECs cultures (42, 43) and in the aortaof the mouse (40, 41, 44), where Notch signalling is uniquelypositioned to attenuate the calcific response to turbulent shearstress within the aortic valves by preventing the expression ofmediators of inflammation and osteogenic markers leading toaortic calcification (41). The role of Notch as mechanosensor inarteries has been recently confirmed by a study showing thatNotch1 activation by shear stress directly regulates vascularbarrier function and it is indispensable for the maintenance ofjunctional integrity, and suppression of proliferation, induced bylaminar shear stress (45, 46). Oppositely to laminar shear stress,dyslipidemia reduces Notch1 in the endothelium leading toincreased hypercholesterolemia-induced atherosclerosis in thedescending aorta (46).

In dyslipidemic mice, expression of Hes5 mediates theproliferation of cells required to repair the endothelium damagedby dyslipidemia (40). Additionally, Notch1 activation in ECs,interferes with NF-kB activity by blocking the transcription ofmiR155 (39, 47) and by inhibiting ERK1/2 activation (48).Taken all together these data suggest that activation of the Notchpathway, possibly through inhibition of ERK/NF-kB signalling,could be responsible for the beneficial effects of ivabradine onthe endothelium in aortic arch of ApoE-deficient mice. The lateincrease of Hes5 protein level concomitant to the endothelialprotection that we found is supportive of the hypothesis.

The hypothesis of a Notch-mediated inhibition of the NF-kBpathway is interesting. This pathway is ‘primed’ for over-activation in atheroprone regions and this ‘priming’ is inducedby disturbed/low shear stress which causes an increase in theexpression of NF-kB components (49). Our analyses show thatearly ivabradine treatment did not alter the expression of NF-kBcomponents, but clearly reduced the expression of several pro-inflammatory genes known to be activated by NF-kB, asdemonstrated by decreased Olr1 gene and protein expressionlevels. It is possible that HR reduction with ivabradine reducesNF-kB activity acting by other mechanisms thandownregulation of transcription of its components. Similarly,our data suggest that ivabradine counteracts AngII-signalling.This is not new, as Custodis et al., showed that in ApoE–/– micefed a high cholesterol diet, treatment with ivabradine for 6weeks, starting at 12 weeks of age, reduces the expression of theAT-1 receptor in the vascular wall (14). We did not observesimilar changes in Agtr1 expression; this discrepancy, likely dueto the difference in diet and treatment duration, suggests, onceagain, that in early stages, HR reduction with ivabradinemodulates other genes involved in the activation of the AngII-pathway. Ang II signalling is activated by dyslipidemicconditions and it requires ERK1/2-mediated activation of NF-kB to induce its downstream genes (50, 51). In agreement withother studies (6), and similarly to other anti-atheroscleroticagents, like modified pullulan (52), ivabradine did not alter the

serum lipid levels, neither we observed changes in theexpression of Ace, but, we did observe a downregulation ofgenes belonging to the MAPK pathway such as Prkca, Rras2,Map3k8 and Mras which, if confirmed by studies on proteinlevels, may suggest that ivabradine may antagonize both AngIIand NF-kB signalling by interfering with the phosphorylationcascade which leads to ERK1/2 activation (50, 51). Thus, it ispossible that early administration of ivabradine results in anintriguing interaction between Notch activation andantagonization of AngII and NF-kB pro-atheroscleroticsignaling (Fig. 10).

There is not conclusive evidence on how ivabradine protectsthe vasculature. Unlike cardiac-specific b-blockers (8, 10, 53),like nebivolol which increases vascular nitric oxide production(54), there is no evidence of a direct effect of ivabradine onendothelium (6, 10). Several studies show that ivabradinedoesn’t affect diastolic function, cardiac output, or bloodpressure both in ApoE–/– and WT mice (7-9). It is worthmentioning that in other animal models, i.e. L-NAME-inducedhypertensive rats, ivabradine, besides lowering HR, is also ableto reduce systolic blood pressure (55). Custodis et al. reportedthat HR reduction with ivabradine improves aortic distensibilityand circumferential cyclic strain in ApoE mice, which might, inturn, modify shear stress pattern (14). More recently, it has beenshown that treatment of low-density lipoprotein receptor(LDLr)-deficient mice with ivabradine alters local mechanicalconditions and improves shear stress conditions in the aorta,without changes in blood pressure, leading to the reduction ofthe expression of inflammatory marker (7). In this study, we didnot investigate the mechanism underlying the effects ofivabradine on the gene expression in the endothelium. In supportwith studies suggesting an effect mediated by shear stressalteration, we report that the expression of many shear stress-regulated genes, including Notch1, was modified by treatment.

The recent clinical trial SIGNIFY (Study Assessing theMorbidity-Mortality Benefits of the If Inhibitor Ivabradine inPatients with Coronary Artery Disease) in patients with CADand preserved left ventricular function, specifically tested thehypothesis that HR reduction with ivabradine reduces theprogression of coronary atherosclerosis and prevents myocardialinfarction and cardiovascular death. Contrary to the previousdata in patients with left ventricular dysfunction and heart failureshowing a clear beneficial outcome from ivabradine treatment(56), and contrary to our experimental findings, in SIGNIFYstudy ivabradine failed to improve patients’ outcome (57).Obviously, it is difficult, if at all possible, to extrapolateexperimental into clinical data but it could be that in CADpatients enrolled in SIGNIFY study, treatment with ivabradinewas applied too late, when atherosclerotic plaques were alreadydeveloped as this was an entry criteria to the trial. Interestingly,recent data on patients with microvascular angina in whomatherosclerosis of the major coronary artery was excluded,ivabradine improved endothelial function and coronary reserve(58-60).

In conclusion, we report that ivabradine treatment, initiatedbefore plaque formation, induces an atheroprotective geneexpression profile and it reduces endothelial damage in the aorticarch of ApoE–/– mice. Potential mechanisms involve activationof Notch pathway and inhibition of ERK/NF-kB.

Acknowledgements: This work was supported by grant fromServier (France) to P.R.

Conflict of interests: Roberto Ferrari reported that hereceived honorarium from Servier for steering committeemembership consulting and speaking, and support for travel tostudy meetings from Servier. In addition, he received personal

48

fees from Boehringer-Ingelheim, Novartis, Merck Serono,Micom, Cipla, and Menarini. Finally, he is a stockholder inMedical Trials Analysis.

REFERENCES

1. Custodis F, Schirmer SH, Baumhakel M, Heusch G, BohmM, Laufs U. Vascular pathophysiology in response toincreased heart rate. J Am Coll Cardiol 2010; 56: 1973-1983.

2. Fox KM, Ferrari R. Heart rate: a forgotten link in coronaryartery disease? Nat Rev Cardiol 2011; 8: 369-379.

3. Kaplan jR, Manuck SB, Clarkson TB. The influence of heartrate on coronary artery atherosclerosis. J CardiovascPharmacol 1987; 10: S100-S103.

4. Perski A, Olsson G, Landou C, de Faire U, Theorell T,Hamsten A. Minimum heart rate and coronaryatherosclerosis: independent relations to global severity andrate of progression of angiographic lesions in men withmyocardial infarction at a young age. Am Heart J 1992; 123:609-616.

5. Beere PA, Glagov S, Zarins CK. Retarding effect of loweredheart rate on coronary atherosclerosis. Science 1984; 226:180-182.

6. Custodis F, Baumhakel M, Schlimmer N, et al. Heart ratereduction by ivabradine reduces oxidative stress, improvesendothelial function, and prevents atherosclerosis inapolipoprotein E-deficient mice. Circulation 2008; 117:2377-2387.

7. Luong L, Duckles H, Schenkel T, et al. Heart rate reductionwith ivabradine promotes shear stress-dependent anti-inflammatory mechanisms in arteries. Thromb Haemost2016; 116: 181-190.

8. Schirmer SH, Degen A, Baumhakel M, et al. Heart-ratereduction by If-channel inhibition with ivabradine restorescollateral artery growth in hypercholesterolemicatherosclerosis. Eur Heart J 2012; 33: 1223-1231.

9. Lauzier B, Vaillant F, Gelinas R, et al. Ivabradine reducesheart rate while preserving metabolic fluxes and energystatus of healthy normoxic working hearts. Am J PhysiolHeart Circ Physiol 2011; 300: H845-H852.

10. Drouin A, Gendron ME, Thorin E, Gillis MA, Mahlberg-Gaudin F, Tardif jC. Chronic heart rate reduction byivabradine prevents endothelial dysfunction indyslipidaemic mice. Br J Pharmacol 2008; 154: 749-757.

11. Bolduc V, Drouin A, Gillis MA, et al. Heart rate-associatedmechanical stress impairs carotid but not cerebral arterycompliance in dyslipidemic atherosclerotic mice. Am JPhysiol Heart Circ Physiol 2011; 301: H2081-H2092.

12. Baumhakel M, Custodis F, Schlimmer N, Laufs U, Bohm M.Heart rate reduction with ivabradine improves erectiledysfunction in parallel to decrease in atherosclerotic plaque loadin ApoE-knockout mice. Atherosclerosis 2010; 212: 55-62.

13. Kroller-Schon S, Schulz E, Wenzel P, et al. Differentialeffects of heart rate reduction with ivabradine in two modelsof endothelial dysfunction and oxidative stress. Basic ResCardiol 2011; 106: 1147-1158.

14. Custodis F, Fries P, Muller A, et al. Heart rate reduction byivabradine improves aortic compliance in apolipoprotein E-deficient mice. J Vasc Res 2012; 49: 432-440.

15. Resnick N, Gimbrone MA. Hemodynamic forces arecomplex regulators of endothelial gene expression. FASEB J1995; 9: 874-882.

16. Foteinos G, Hu Y, Xiao Q, Metzler B, Xu Q. Rapidendothelial turnover in atherosclerosis-prone areas coincideswith stem cell repair in apolipoprotein E-deficient mice.Circulation 2008; 117: 1856-1863.

17. Sakao S, Taraseviciene-Stewart L, Lee jD, Wood K, CoolCD, Voelkel NF. Initial apoptosis is followed by increasedproliferation of apoptosis-resistant endothelial cells. FASEBJ 2005; 19: 1178-1180.

18. Krenek P, Hamaide MC, Morel N, Wibo M. A simplemethod for rapid separation of endothelial and smoothmuscle mRNA reveals Na/K+-ATPase alpha-subunitdistribution in rat arteries. J Vasc Res 2006; 43: 502-510.

19. Huang DW, Sherman BT, Lempicki RA. Systematic andintegrative analysis of large gene lists using DAVIDbioinformatics resources. Nat Protoc 2009; 4: 44-57.

20. Caliceti C, Aquila G, Pannella M, et al. 17beta-estradiolenhances signalling mediated by VEGF-A-delta-like ligand4-notch1 axis in human endothelial cells. PLoS One 2013; 8:e71440. doi: 10.1371/journal.pone.0071440

21. Fortini F, Vieceli Dalla SF, Caliceti C, et al. Estrogenreceptor beta-dependent Notch1 activation protects vascularendothelium against tumor necrosis factor alpha(TNFalpha)-induced apoptosis. J Biol Chem 2017; 292:18178-18191.

22. Iiyama K, Hajra L, Iiyama M, et al. Patterns of vascular celladhesion molecule-1 and intercellular adhesion molecule-1expression in rabbit and mouse atherosclerotic lesions and atsites predisposed to lesion formation. Circ Res 1999; 85:199-207.

23. Nakashima Y, Plump AS, Raines EW, Breslow jL, Ross R.ApoE-deficient mice develop lesions of all phases ofatherosclerosis throughout the arterial tree. ArteriosclerThromb 1994; 14: 133-140.

24. Pritchard KA, Schwarz SM, Medow MS, Stemerman MB.Effect of low-density lipoprotein on endothelial cellmembrane fluidity and mononuclear cell attachment. Am JPhysiol 1991; 260: C43-C49.

25. Diczfalusy U, Olofsson KE, Carlsson AM, et al. Markedupregulation of cholesterol 25-hydroxylase expression bylipopolysaccharide. J Lipid Res 2009; 50: 2258-2264.

26. Gold ES, Ramsey SA, Sartain Mj, et al. ATF3 protects againstatherosclerosis by suppressing 25-hydroxycholesterol-inducedlipid body formation. J Exp Med 2012; 209: 807-817.

27. Liu Y, Chen BP, Lu M, et al. Shear stress activation ofSREBP1 in endothelial cells is mediated by integrins.Arterioscler Thromb Vasc Biol 2002; 22: 76-81.

28. Gargalovic PS, Gharavi NM, Clark Mj, et al. The unfoldedprotein response is an important regulator of inflammatorygenes in endothelial cells. Arterioscler Thromb Vasc Biol2006; 26: 2490-2496.

29. Kallin A, johannessen LE, Cani PD, et al. SREBP-1regulates the expression of heme oxygenase 1 and thephosphatidylinositol-3 kinase regulatory subunit p55gamma. J Lipid Res 2007; 48: 1628-1636.

30. Luo W, Bodary PF, Shen Y, et al. Leptin receptor-inducedSTAT3-independent signaling pathways are protectiveagainst atherosclerosis in a murine model of obesity andhyperlipidemia. Atherosclerosis 2011; 214: 81-85.

31. Chatzizisis YS, Coskun AU, jonas M, Edelman ER,Feldman CL, Stone PH. Role of endothelial shear stress inthe natural history of coronary atherosclerosis and vascularremodeling: molecular, cellular, and vascular behavior. J AmColl Cardiol 2007; 49: 2379-2393.

32. Benedito R, Roca C, Sorensen I, et al. The Notch ligandsDll4 and jagged1 have opposing effects on angiogenesis.Cell 2009; 137: 1124-1135.

33. Simmons Mj, Serra R, Hermance N, Kelliher MA.NOTCH1 inhibition in vivo results in mammary tumorregression and reduced mammary tumorsphere-formingactivity in vitro. Breast Cancer Res 2012; 14: R126. doi:10.1186/bcr3321

49

34. Lu TM, Luo Yj, Yu jK. BMP and Delta/Notch signalingcontrol the development of amphioxus epidermal sensoryneurons: insights into the evolution of the peripheral sensorysystem. Development 2012; 139: 2020-2030.

35. Nus M, Martinez-Poveda B, MacGrogan D. Endothelialjag1-RBPj signalling promotes inflammatory leucocyterecruitment and atherosclerosis. Cardiovasc Res 2016; Aug.5: cvw193. Epub ahead of print.

36. Liu Zj, Tan Y, Beecham GW, et al. Notch activation inducesendothelial cell senescence and pro-inflammatory response:implication of Notch signaling in atherosclerosis.Atherosclerosis 2012; 225: 296-303.

37. Briot A, Civelek M, Seki A, et al. Endothelial NOTCH1 issuppressed by circulating lipids and antagonizesinflammation during atherosclerosis. J Exp Med 2015; 212:2147-2163.

38. Quillard T, Charreau B. Impact of notch signaling oninflammatory responses in cardiovascular disorders. Int JMol Sci 2013; 14: 6863-6888.

39. Wang L, Zhang H, Rodriguez S, et al. Notch-dependentrepression of miR-155 in the bone marrow niche regulateshematopoiesis in an NF-kappaB-dependent manner. CellStem Cell 2014; 15: 51-65.

40. Schober A, Nazari-jahantigh M, Wei Y, et al. MicroRNA-126-5p promotes endothelial proliferation and limitsatherosclerosis by suppressing Dlk1. Nat Med 2014; 20:368-376.

41. Theodoris CV, Li M, White MP, et al. Human diseasemodeling reveals integrated transcriptional and epigeneticmechanisms of NOTCH1 haploinsufficiency. Cell 2015;160: 1072-1086.

42. Masumura T, Yamamoto K, Shimizu N, Obi S, Ando j. Shearstress increases expression of the arterial endothelial markerephrinB2 in murine ES cells via the VEGF-Notch signalingpathways. Arterioscler Thromb Vasc Biol 2009; 29: 2125-2131.

43. Walshe TE, Connell P, Cryan L, et al. Microvascular retinalendothelial and pericyte cell apoptosis in vitro: role ofHedgehog and Notch signaling. Invest Ophthalmol Vis Sci2011; 52: 4472-4483.

44. Morelli MB, Aquila G, Bonora M, et al. Differentialexpression of Notch pathway components in atheroprotectedversus atherosusceptible regions of endothelium of mouseaorta. [abstract]. Eur Heart J 2014; 35 (Suppl.): 704.

45. Polacheck Wj, Kutys ML, Yang j, et al. A non-canonicalNotch complex regulates adherens junctions and vascularbarrier function. Nature 2017; 552: 258-262.

46. Mack jj, Mosqueiro TS, Archer Bj, et al. NOTCH1 is amechanosensor in adult arteries. Nat Commun 2017; 8:1620. doi: 10.1038/s41467-017-01741-8

47. Lee KS, Kim j, Kwak SN, et al. Functional role of NF-kappaB in expression of human endothelial nitric oxidesynthase. Biochem Biophys Res Commun 2014; 448: 101-107.

48. Liu Zj, Xiao M, Balint K, et al. Inhibition of endothelial cellproliferation by Notch1 signaling is mediated by repressingMAPK and PI3K/Akt pathways and requires MAML1.FASEB J 2006; 20: 1009-1011.

49. Hajra L, Evans AI, Chen M, Hyduk Sj, Collins T, CybulskyMI. The NF-kappaB signal transduction pathway in aorticendothelial cells is primed for activation in regionspredisposed to atherosclerotic lesion formation. Proc NatlAcad Sci USA 2000; 97: 9052-9057.

50. Alonso F, Krattinger N, Mazzolai L, et al. An angiotensin II-and NF-kappaB-dependent mechanism increases connexin43 in murine arteries targeted by renin-dependenthypertension. Cardiovasc Res 2010; 87: 166-176.

51. Indra MR, Karyono S, Ratnawati R, Malik SG. Quercetinsuppresses inflammation by reducing ERK1/2phosphorylation and NF kappa B activation in leptin-induced human umbilical vein endothelial cells (HUVECs).BMC Res Notes 2013; 6: 275. doi: 10.1186/1756-0500-6-275

52. Stefan j, Kus K, Wisniewska A, et al. New cationicallymodified pullulan attenuates atherogenesis and influenceslipid metabolism in apoE-knockout mice. J PhysiolPharmacol 2016; 67: 739-749.

53. Garlichs CD, Zhang H, Mugge A, Daniel WG. Beta-blockersreduce the release and synthesis of endothelin-1 in humanendothelial cells. Eur J Clin Invest 1999; 29: 12-16.

54. Kus K, Wisniewska A, Toton-Zuranska j, Olszanecki R,jawien j, Korbut R. Significant deterioration of anti-atherogenic efficacy of nebivolol in a double (apolipoprotein Eand endothelial nitric oxide synthase) knockout mouse modelof atherosclerosis in comparison to single (apolipoprotein E)knockout model. J Physiol Pharmacol 2014; 65: 877-881.

55. Aziriova S, Repova K, Krajcirovicova K, et al. Effect ofivabradine, captopril and melatonin on the behaviour of rats inL-nitro-arginine methyl ester-induced hypertension. J PhysiolPharmacol 2016; 67: 895-902.

56. Swedberg K, Komajda M, Bohm M, et al. Ivabradine andoutcomes in chronic heart failure (SHIFT): a randomisedplacebo-controlled study. Lancet 2010; 376: 875-885.

57. Fox K, Ford I, Steg PG, Tardif jC, Tendera M, Ferrari R.Ivabradine in stable coronary artery disease without clinicalheart failure. N Engl J Med 2014; 371: 1091-1099.

58. Skalidis EI, Hamilos MI, Chlouverakis G, Zacharis EA,Vardas PE. Ivabradine improves coronary flow reserve inpatients with stable coronary artery disease. Atherosclerosis2011; 215: 160-165.

59. van der Hoeven NW, van Royen N. The effect of heart ratereduction by ivabradine on collateral function in patientswith chronic stable coronary artery disease, another funnyaspect of the funny channel? Heart 2014; 100: 98-99.

60. Gloekler S, Traupe T, Stoller M, et al. The effect of heart ratereduction by ivabradine on collateral function in patientswith chronic stable coronary artery disease. Heart 2014;100: 160-166.

61. Passino C, Del Ry S, Severino S, et al. C-type natriureticpeptide expression in patients with chronic heart failure:effects of aerobic training. Eur J Cardiovasc Prev Rehabil2008; 15: 168-172.

62. Chen G, Zhao j, Yin Y, et al. C-type natriuretic peptideattenuates LPS-induced endothelial activation: involvementof p38, Akt, and NF-kappaB pathways. Amino Acids 2014;46: 2653-2663.

63. Zhu W, Saddar S, Seetharam D, et al. The scavenger receptorclass B type I adaptor protein PDZK1 maintains endothelialmonolayer integrity. Circ Res 2008; 102: 480-487.

64. Turner EC, Mulvaney EP, Reid HM, Kinsella BT. Interactionof the human prostacyclin receptor with the PDZ adapterprotein PDZK1: role in endothelial cell migration andangiogenesis. Mol Biol Cell 2011; 22: 2664-2679.

65. Al-Aly Z, Shao jS, Lai CF, et al. Aortic Msx2-Wntcalcification cascade is regulated by TNF-alpha-dependentsignals in diabetic Ldlr–/– mice. Arterioscler Thromb VascBiol 2007; 27: 2589-2596.

66. Li Z, Shi HY, Zhang M. Targeted expression of maspin intumor vasculatures induces endothelial cell apoptosis.Oncogene 2005; 24: 2008-2019.

67. Salter RC, Ashlin TG, Kwan AP, Ramji DP. ADAMTSproteases: key roles in atherosclerosis? J Mol Med (Berl)2010; 88: 1203-1211.

68. Yuan D, Sun G, Zhang R, et al. Connexin 43 expressed inendothelial cells modulates monocyteendothelial adhesion

50

by regulating cell adhesion proteins. Mol Med Rep 2015; 12:7146-7152.

69. Huang j, Stohl LL, Zhou X, Ding W, Granstein RD.Calcitonin gene-related peptide inhibits chemokineproduction by human dermal microvascular endothelialcells. Brain Behav Immun 2011; 25: 787-799.

70. Hernanz A, De Miguel E, Romera N, Perez-Ayala C, Gijonj, Arnalich F. Calcitonin gene-related peptide II, substance Pand vasoactive intestinal peptide in plasma and synovialfluid from patients with inflammatory joint disease. Br JRheumatol 1993; 32: 31-35.

71. Savas S, Hyde A, Stuckless SN, Parfrey P, Younghusband HB,Green R. Serotonin transporter gene (SLC6A4) variations areassociated with poor survival in colorectal cancer patients.PLoS One 2012; 7: e38953. doi: 10.1371/journal.pone.0038953

72. Morawietz H, Duerrschmidt N, Niemann B, Galle j,Sawamura T, Holtz j. Induction of the oxLDL receptorLOX-1 by endothelin-1 in human endothelial cells. BiochemBiophys Res Commun 2001; 284: 961-965.

73. Bond AR, Hultgardh-Nilsson A, Knutsson A, jackson CL,Rauch U. Cartilage oligomeric matrix protein (COMP) inmurine brachiocephalic and carotid atherosclerotic lesions.Atherosclerosis 2014; 236: 366-372.

74. Torban E, Eccles MR, Favor j, Goodyer PR. PAX2suppresses apoptosis in renal collecting duct cells. Am JPathol 2000; 157: 833-842.

75. Kim Wj, Bae EM, Kang Yj, et al. Glucocorticoid-inducedtumour necrosis factor receptor family related protein(GITR) mediates inflammatory activation of macrophagesthat can destabilize atherosclerotic plaques. Immunology2006; 119: 421-429.

76. Marzoll A, Nagy N, Wordehoff L, et al. Cyclooxygenaseinhibitors repress vascular hyaluronan-synthesis in murineatherosclerosis and neointimal thickening. J Cell Mol Med2009; 13: 3713-3719.

77. Lin j, Chang W, Dong j, et al. Thymic stromallymphopoietin over-expressed in human atherosclerosis:potential role in Th17 differentiation. Cell Physiol Biochem2013; 31: 305-318.

78. jia M, Dahlman-Wright K, Gustafsson jA. Estrogen receptoralpha and beta in health and disease. Best Pract Res ClinEndocrinol Metab 2015; 29: 557-568.

79. Hassan HH, Denis M, Krimbou L, Marcil M, Genest j.Cellular cholesterol homeostasis in vascular endothelialcells. Can J Cardiol 2006; 22 (Suppl. B): 35B-40B.

80. Burleigh ME, Babaev VR, Yancey PG, et al.Cyclooxygenase-2 promotes early atherosclerotic lesionformation in ApoE-deficient and C57BL/6 mice. J Mol CellCardiol 2005; 39: 443-452.

81. Bhayadia R, Schmidt BM, Melk A, Homme M. Senescence-induced oxidative stress causes endothelial dysfunction. J Gerontol A Biol Sci Med Sci 2016; 71: 161-169.

82. Lewis RD, Perry Mj, Guschina IA, jackson CL, Morgan BP,Hughes TR. CD55 deficiency protects againstatherosclerosis in ApoE-deficient mice via C3a modulationof lipid metabolism. Am J Pathol 2011; 179: 1601-1607.

83. Cox MA, jackson j, Stanton M, et al. Short-chain fatty acidsact as antiinflammatory mediators by regulatingprostaglandin E(2) and cytokines. World J Gastroenterol2009; 15: 5549-5557.

84. Garcia-Heredia A, Marsillach j, Rull A, et al. Paraoxonase-1 inhibits oxidized low-density lipoprotein-inducedmetabolic alterations and apoptosis in endothelial cells: anondirected metabolomic study. Mediators Inflamm 2013;2013: 156053. doi: 10.1155/2013/156053

85. Charan RA, johnson BN, Zaganelli S, Nardozzi jD, LaVoieMj. Inhibition of apoptotic Bax translocation to the

mitochondria is a central function of parkin. Cell Death Dis2014; 5: e1313. doi: 10.1038/cddis.2014.278

86. Sandoval R, Malik AB, Minshall RD, Kouklis P, Ellis CA,Tiruppathi C. Ca(2+) signalling and PKCalpha activateincreased endothelial permeability by disassembly of VE-cadherin junctions. J Physiol 2001; 533: 433-445.

87. DeBose-Boyd RA. Feedback regulation of cholesterolsynthesis: sterol-accelerated ubiquitination and degradationof HMG CoA reductase. Cell Res 20018; 18: 609-621.

88. Lorbek G, Lewinska M, Rozman D. Cytochrome P450s inthe synthesis of cholesterol and bile acids - from mousemodels to human diseases. FEBS J 2012; 279: 1516-1533.

89. Shih DQ, Bussen M, Sehayek E, et al. Hepatocyte nuclearfactor-1alpha is an essential regulator of bile acid and plasmacholesterol metabolism. Nat Genet 2001; 27: 375-382.

90. O’Rourke L, Yeaman Sj, Shepherd PR. Insulin and leptinacutely regulate cholesterol ester metabolism inmacrophages by novel signaling pathways. Diabetes 2001;50: 955-961.

91. Liu XY, Dangel AW, Kelley RI, et al. The gene mutated in barepatches and striated mice encodes a novel 3beta-hydroxysteroid dehydrogenase. Nat Genet 1999; 22: 182-187.

92. Boussicault L, Alves S, Lamaziere A, et al. CYP46A1, therate-limiting enzyme for cholesterol degradation, isneuroprotective in Huntington’s disease. Brain 2016; 139:953-970.

93. Lund EG, Kerr TA, Sakai j, Li WP, Russell DW. cDNAcloning of mouse and human cholesterol 25-hydroxylases,polytopic membrane proteins that synthesize a potentoxysterol regulator of lipid metabolism. J Biol Chem 1998;273: 34316-34327.

94. Luo M, Chodisetti G, Riederer B, Tian DA, Yeruva S,Seidler U. Interleukin-1b mediated transcriptionaldownregulation of NHERF family PDZ adaptor proteinPDZK1 in CACO-2BBE cells. Acta Physiol 2013; 207(Suppl. 694): P163.

95. Qin S, Taglienti M, Cai L, Zhou j, Kreidberg jA. c-Met andNF-kappaB-dependent overexpression of Wnt7a and -7b andPax2 promotes cystogenesis in polycystic kidney disease. J AmSoc Nephrol 2012; 23: 1309-1318.

96. Guo F, Kang S, Zhou P, Guo L, Ma L, Hou j. Maspinexpression is regulated by the non-canonical NF-kappaBsubunit in androgen-insensitive prostate cancer cell lines.Mol Immunol 2011; 49: 8-17.

97. Gabriel HD, Strobl B, Hellmann P, Buettner R, WinterhagerE. Organization and regulation of the rat Cx31 gene.Implication for a crucial role of the intron region. Eur JBiochem 2001; 268: 1749-1759.

98. Bowen Ej, Schmidt TW, Firm CS, Russo AF, Durham PL.Tumor necrosis factor-alpha stimulation of calcitonin gene-related peptide expression and secretion from rat trigeminalganglion neurons. J Neurochem 2006; 96: 65-77.

99. Flattem NL, Blakely RD. Modified structure of the humanserotonin transporter promoter. Mol Psychiatry 2000; 5:110-115.

100. Nagase M, Abe j, Takahashi K, Ando j, Hirose S, Fujita T.Genomic organization and regulation of expression of thelectin-like oxidized low-density lipoprotein receptor (LOX-1) gene. J Biol Chem 1998; 273: 33702-33707.

101. Bitko V, Velazquez A, Yang L, Yang YC, Barik S.Transcriptional induction of multiple cytokines by humanrespiratory syncytial virus requires activation of NF-kappaB and is inhibited by sodium salicylate and aspirin.Virology 1997; 232: 369-378.

102. Lai Y, Bai X, Zhao Y, et al. ADAMTS-7 forms a positivefeedback loop with TNF-alpha in the pathogenesis ofosteoarthritis. Ann Rheum Dis 2014; 73: 1575-1584.

51

103. Saavalainen K, Tammi MI, Bowen T, Schmitz ML,Carlberg C. Integration of the activation of the humanhyaluronan synthase 2 gene promoter by common cofactorsof the transcription factors retinoic acid receptor andnuclear factor kappaB. J Biol Chem 2007; 282: 11530-11539.

104. Ackerman WE, Summerfield TL, Vandre DD, RobinsonjM, Kniss DA. Nuclear factor-kappa B regulates inducibleprostaglandin E synthase expression in human amnionmesenchymal cells. Biol Reprod 2008; 78: 68-76.

105. Izban MG, Nowicki Bj, Nowicki S. 1,25-Dihydroxyvitamin D3 promotes a sustained LPS-inducedNF-kappaB-dependent expression of CD55 in humanmonocytic THP-1 cells. PLoS One 2012; 7: e49318. doi:10.1371/journal.pone.0049318

106. Zhang M, Zhou SH, Zhao SP, Liu QM, Li XP, Shen XQ.Irbesartan attenuates Ang II-induced BMP-2 expression inhuman umbilical vein endothelial cells. Vasc Med 2008; 13:239-245.

107. Morawietz H, Rueckschloss U, Niemann B, et al.Angiotensin II induces LOX-1, the human endothelialreceptor for oxidized low-density lipoprotein. Circulation1999; 100: 899-902.

108. Sendra j, Llorente-Cortes V, Costales P, Huesca-Gomez C,Badimon L. Angiotensin II upregulates LDL receptor-related protein (LRP1) expression in the vascular wall: anew pro-atherogenic mechanism of hypertension.Cardiovasc Res 2008; 78: 581-589.

109. Romero DG, Plonczynski M, Vergara GR, Gomez-SanchezEP, Gomez-Sanchez CE. Angiotensin II early regulatedgenes in H295R human adrenocortical cells. PhysiolGenomics 2004; 19: 106-116.

110. Zhao H, Li M, Wang L, et al. Angiotensin II induces TSLPvia an AT1 receptor/NF-KappaB pathway, promoting Th17differentiation. Cell Physiol Biochem 2012; 30: 1383-1397.

111. Schlondorff D, DeCandido S, Satriano jA. Angiotensin IIstimulates phospholipases C and A2 in cultured ratmesangial cells. Am J Physiol 1987; 253: C113-C120.

112. Takata Y, Liu j, Yin F, et al. PPARdelta-mediatedantiinflammatory mechanisms inhibit angiotensin II-accelerated atherosclerosis. Proc Natl Acad Sci USA 2008;105: 4277-4282.

113. Iwaya S, Oikawa M, Chen Y, Takeishi Y. Phosphodiesterase3A1 protects the heart against angiotensin II-induced cardiacremodeling through regulation of transforming growthfactor-beta expression. Int Heart J 2014; 55: 165-168.

114. Esteban V, Mendez-Barbero N, jimenez-Borreguero Lj, etal. Regulator of calcineurin 1 mediates pathological vascularwall remodeling. J Exp Med 2011; 208: 2125-2139.

115. Zhang Z, Xiao Z, Diamond SL. Shear stress induction of C-type natriuretic peptide (CNP) in endothelial cells isindependent of NO autocrine signaling. Ann Biomed Eng1999; 27: 419-426.

116. Avvisato CL, Yang X, Shah S, et al. Mechanical forcemodulates global gene expression and beta-cateninsignaling in colon cancer cells. J Cell Sci 2007; 120:2672-2682.

117. Ding Z, Liu S, Deng X, Fan Y, Wang X, Mehta jL.Hemodynamic shear stress modulates endothelial cellautophagy: role of LOX-1. Int J Cardiol 2015; 184: 86-95.

118. Brooks AR, Lelkes PI, Rubanyi GM. Gene expressionprofiling of vascular endothelial cells exposed to fluidmechanical forces: relevance for focal susceptibility toatherosclerosis. Endothelium 2004; 11: 45-57.

119. Dolan jM, Sim Fj, Meng H, Kolega j. Endothelial cellsexpress a unique transcriptional profile under very highwall shear stress known to induce expansive arterialremodeling. Am J Physiol Cell Physiol 2012; 302: C1109-C1118.

120. Vion AC, Ramkhelawon B, Loyer X, et al. Shear stressregulates endothelial microparticle release. Circ Res 2013;112: 1323-1333.

121. Passerini AG, Polacek DC, Shi C, et al. Coexistingproinflammatory and antioxidative endothelialtranscription profiles in a disturbed flow region of theadult porcine aorta. Proc Natl Acad Sci USA 2004; 101:2482-2487.

122. Peters DG, Zhang XC, Benos PV, Heidrich-O’Hare E,Ferrell RE. Genomic analysis of immediate/early responseto shear stress in human coronary artery endothelial cells.Physiol Genomics 2002; 12: 25-33.

123. White Sj, Hayes EM, Lehoux S, jeremy jY, Horrevoets Aj,Newby AC. Characterization of the differential response ofendothelial cells exposed to normal and elevated laminarshear stress. J Cell Physiol 2011; 226: 2841-2848.

124. Andersson M, Karlsson L, Svensson PA, et al. Differentialglobal gene expression response patterns of humanendothelium exposed to shear stress and intraluminalpressure. J Vasc Res 2005; 42: 441-452.

125. Chiu jj, Lee PL, Chang SF, et al. Shear stress regulatesgene expression in vascular endothelial cells in response totumor necrosis factor-alpha: a study of the transcriptionprofile with complementary DNA microarray. J Biomed Sci2005; 12: 481-502.

126. Ni CW, Wang DL, Lien SC, Cheng jj, Chao Yj, Hsieh Hj.Activation of PKC-epsilon and ERK1/2 participates in shear-induced endothelial MCP-1 expression that is repressed bynitric oxide. J Cell Physiol 2003; 195: 428-434.

127. Glossop jR, Cartmell SH. Effect of fluid flow-inducedshear stress on human mesenchymal stem cells: differentialgene expression of IL1B and MAP3K8 in MAPKsignaling. Gene Expr Patterns 2009; 9: 381-388.

128. Lenz M, Muller Fj, Zenke M, Schuppert A. Principalcomponents analysis and the reported low intrinsicdimensionality of gene expression microarray data. Sci Rep2016; 6: 25696. doi: 10.1038/srep25696

R e c e i v e d : August 1, 2017A c c e p t e d : February 15, 2018

Author’s address: Dr. Paola Rizzo, Department ofMorphology, Surgery and Experimental Medicine; University ofFerrara, 64/B via Fossato di Mortara, 44121 Ferrara, Italy.E-mail: rzzpla@unife.it

52